I would approach the question of a team by asking what values we have as Debian science developers. I have already seen Andreas representing consistency and attention to detail, and I would agree that we should display these if we have a new team. I have some other aesthetics too:
When I first got into scientific programming about twenty years ago, I wanted to be as standard as possible. At first, that meant using Numerical Recipes. Next, it seemed to mean using ANSI C/gcc. Then, the Gnu Scientific Library. After a long time learning, I thought it was a good goal to eventually become part of the GSL. But that took years and GSL continued to grow while I refactored my programs. Now, GSL is so big, I think maybe it does not want any more big software modules. Therefore, I decided to fall back to GObject/glib which is still pretty generic, with standard C and C++ mostly. And I am just 100% compatible and dependent on GSL, mostly for vector and matrix types. I now have about a dozen scientific software packages and am happy to organize them wherever it is convenient. But, I enjoy using Mercurial more than Subversion, so I hope that is possible. I enjoy making Erlang-style multi-language connectors (ncd --server) for my programs and am excited to do that in 2008. Just explain how the debian-science project can benefit my packages and I am sure my packages will move themselves onto the project. As you can tell from my focus, I like to explain the possibility for team cooperation based on development principles. If we have overlap, there is probably a reason to watch each other's repositories more carefully and maybe work more on each other's sourcecodes. If we have little overlap, this may be a frustrating and ultimately futile or pointless distraction. For 2008, my focus is on testing, documenting, and refactoring I think, as well as learning the Debian rules better. Best regards, Rudi On Jan 28, 2008 12:29 PM, Sylvestre Ledru <[EMAIL PROTECTED]> wrote: > Le lundi 28 janvier 2008 à 15:13 -0500, Ross Boylan a écrit : > > On Sun, 2008-01-27 at 06:38 +0100, Christophe Prud'homme wrote: > > > in my opinion the scope is quite broad : it spans the "sciences of > > > computing", > > > So pkg-scicomp should definitely not host all possible "science" > > > packages. > > > > Speaking only as an onlooker, I thought the original idea (for > > pkg-science) was to make the group for all (packaging of) > > science-related stuff. "sciences of computing" sounds like "computer > > science" to me, which is quite a bit narrower. I think you have > > something different in mind than either "all science" or "computer > > science," but I'm not sure what. > Actually, it was the original idea. However, to work, we need a > "critical mass" to work... And this mass is already existing on scicomp. > > If we are confident that both will exist and work, let's go! > > Sylvestre > > -- Which is worse, ignorance or apathy? Who knows? Who cares? --Erich Schubert
